Vol. 421: 173–182, 2011 MARINE ECOLOGY PROGRESS SERIES Published January 17 doi: 10.3354/meps08895 Mar Ecol Prog Ser

Genetic parentage in the squat lobsters Munida rugosa and M. sarsi (Crustacea, Anomura, Galatheidae)

Deborah A. Bailie, Rosaleen Hynes, Paulo A. Prodöhl*

School of Biological Sciences, Queen’s University Belfast, Medical Biology Centre, 97 Lisburn Road, Belfast BT9 7BL, UK

ABSTRACT: Munida is the most diverse and cosmopolitan genus of the galatheid squat lobsters. The group has attracted much attention in recent years from both systematic and evolutionary perspec- tives, yet information on the biology, ecology and evolution of this genus is very limited. We investi- gated the genetic parentage of 2 North Atlantic species, M. rugosa and M. sarsi, sampled from the Clyde Sea on the west coast of Scotland. Microsatellite markers were used to establish the parental contribution from embryos of berried females (M. rugosa, n = 25 and M. sarsi, n = 5). The frequency of multiple paternity observed in both species (86% for M. rugosa and 100% for M. sarsi) is the high- est ever reported for any marine crustaceans. Invariably more than 2 sires were involved in each case (minimum of 2 to 3 for M. rugosa and 4 for M. sarsi). Our findings indicate that multiple paternity is likely to be the norm in both species. Within most multiply sired broods, sire contribution was highly skewed towards a single male (66% of broods for M. rugosa and 100% for M. sarsi). Furthermore, embryos from different sires were randomly distributed across the female’s brood patch. This is the first report of multiple paternity in galatheids. While a number of theories can account for the high incidence of multiple paternity in these species (e.g. convenience polyandry as a result of cryptic female choice, forced copulations, the influence of fishing pressures), at present it is not possible to disentangle their individual and/or combined effects.

KEY WORDS: Galatheids · Crustacean mating system · Polyandry · Munida spp.

Resale or republication not permitted without written consent of the publisher

INTRODUCTION (Sainte-Marie et al. 1997), convenience polyandry (Thiel & Hinojosa 2003) and cryptic female choice The mating behaviour of decapod crustaceans has (Walker et al. 2002, Thiel & Hinojosa 2003). received considerable attention in recent years (re- A major difficulty in examining the mating behav- viewed by Correa & Thiel 2003, Duffy & Thiel 2007). iour of marine crustaceans in general is associated with Many studies have elaborated on male mating behav- the obvious logistical difficulties of observing mating iours such as mate guarding and aggression (Wada et interactions in their natural environment. The elusive al. 1997, Jivoff & Hines 1998, Rondeau & Sainte-Marie burrowing behaviour and/or nocturnal activities of 2001), or sperm competition. The latter has focused on many species (De Grave & Turner 1997) further com- sperm stratification (Urbani et al. 1998), removal plicate studies on mating. While laboratory experi- (Beninger et al. 1991) and/or limitation (MacDiarmid & ments have contributed to a better understanding of Butler 1999, Rondeau & Sainte-Marie 2001, Rubolini et the mating behaviour of some species (e.g. Urbani et al. 2005). More recently, attention has shifted towards al. 1998, Thiel & Hinojosa 2003), the mating systems of the examination of female behaviour. These studies the majority of crustaceans are still poorly understood. have addressed questions related to mate selection Molecular studies are now providing a powerful alter-

*Corresponding author. Email: [email protected] © Inter-Research 2011 · www.int-res.com 174 Mar Ecol Prog Ser 421: 173–182, 2011

native for investigating the mating strategies of other- cephalothorax, forming a spawning chamber where wise elusive and/or difficult target species (e.g. eggs are brooded until they hatch. It is not known Bilodeau et al. 2005) whether a female can mate with more than 1 male dur- Molecular investigations of paternity among marine ing a breeding season. While the timing of reproduc- crustaceans are still rare in comparison with other tion is unclear for M. sarsi, in western Scotland M. taxa. With a few exceptions, focusing on brachyuran rugosa produce broods from November, with hatching crabs (Urbani et al. 1998, McKeown & Shaw 2008), the occurring from March to May (Lebour 1930, Zainal majority of other studies on marine crustaceans have 1990, Coombes 2002). Extruded broods contain up to reported on the incidence of multiple paternity, e.g. ~32 000 eggs which are approximately 0.5 mm in porcelain crab Petrolisthes cinctipes (Toonen 2004), diameter (Wenner 1982, Tapella et al. 2002). crayfish Orconectes placidus (Walker et al. 2002), Other crustacean species displaying mating behav- American lobster Homarus americanus (Gosselin et iour similar to squat lobsters, including forced copula- al. 2005) and ghost shrimp Callichirus islagrande tion (e.g. crayfish) and lack of sperm storage structures (Bilodeau et al. 2005). Knowledge of mating strategies (e.g. rock shrimp), are characterised by multiple mat- is particularly relevant for species subject to intense ing. While it is still unknown as to whether multiple fishing exploitation. For instance, in the American lob- mating behaviour translates into multiple genetic ster, Gosselin et al. (2005) showed that mating strate- paternity in rock shrimp, this has been confirmed for gies can vary geographically in correlation with differ- crayfish (Walker et al. 2002). Thus, we hypothesised ent levels of exploitation. The authors suggested that that genetic multiple paternity is the mating strategy of selective fishing pressure for large males could be the galatheid species investigated. In order to test this responsible for such a pattern. hypothesis in Munida rugosa and M. sarsi, we gener- Munida is the most diverse and cosmopolitan genus ated and compared microsatellite multilocus geno- of the galatheid squat lobsters and, as such, has types of females and their egg broods in order to attracted much attention in recent years from both a deduce paternal contribution. systematic and evolutionary perspective. However, information regarding the biology, ecology and evolu- tion of this genus is still limited. Munida rugosa and M. MATERIALS AND METHODS sarsi are endemic to the north-eastern Atlantic. They inhabit rocky or soft mud substrata in shelf waters Study sites and collections. Given their elusive ranging from the shallow to the deeper continental nature and habitat preferences (individuals often bury slope. Similar to other Munida species, M. rugosa and themselves in muddy substrata), it is notoriously diffi- M. sarsi are gonochoric, but little is known about the cult to study the mating behaviour of these species in mating behaviour of these species. They are benthic as their natural environment. This is particularly the case adults with a planktotrophic larval phase and a devel- for egg-carrying females, making them very hard to opmental period that lasts 3 to 4 mo (Schmidt 1965, obtain. Indeed, at least for Munida rugosa, fisheries Gore 1979, Van Dover & Williams 1991). Clarification have been reported to be biased towards males of the mating system of Munida species is now of par- (Coombes 2002). Furthermore, in comparison to M. ticular relevance not only because of the increasing rugosa, M. sarsi, as the deeper water species, is consid- target fisheries, but also due to the impact from indi- erably more difficult to sample. Thus, despite consider- rect fishing pressures. In western Scotland, for able effort, sampling success for ovigerous females instance, M. rugosa comprises a large component of was limited. M. rugosa (n = 25) and M. sarsi (n = 5) bycatch from the large-scale commercial Nephrops ovigerous females were collected by a 2 m beam trawl norvegicus fisheries (Bergmann & Moore 2001). using a 50 mm mesh in the Clyde Sea (western Scot- Pothanikat (2005) produced the only report detailing land) in March 2005 and February 2006. The abdomen the mating behaviour of Munida sarsi, which was and associated eggs from each female were stored in based upon laboratory observations. The report stated 99% molecular grade ethanol for subsequent analysis. that coerced mating occurs in the hard-shelled state, Species identification was confirmed by genetic with a male holding a female in place with his che- screening using diagnostic mtDNA markers (Bailie lipeds, while trying to insert a spermatophore into the 2008). abdominal area of the female with the help of his fifth DNA extraction and amplification of microsatel- pereiopods. Females possess no internal organs for lites. Genomic DNA was extracted from the muscle tis- sperm storage, and there is no information as to how sue of the ovigerous females as described by Taggart the female handles the spermatophore or how egg et al. (1992). For Munida rugosa samples, 11 eggs were insemination takes place. Following insemination, randomly collected from each of the 4 pleopodal however, the female’s abdomen is cupped under the regions (totalling 44 eggs female–1). A similar sampling Bailie et al.: Multiple paternity in Munida spp. 175

strategy was carried out for M. sarsi, except that 23 loaded into a 25 cm 6% 1× TBE polyacrylamide gel us- eggs were randomly removed from each of the 3 ing a commercially available size-standard ladder for the pleopodal regions (totalling 69 eggs female–1). This Li-Cor system (Microstep-13a/b and 20a, Microzone) to sequential egg sampling strategy (i.e. obtaining sam- accurately estimate the size of allelic fragments. Gels ples from all pleopodal regions) was carried out in were run on the Li-Cor system at a constant power of order to account for any potential variation in the tim- 40 W at a temperature of ~50°C for 1 to 2 h. Genotypic ing or order of egg extrusion by females. DNA egg scoring was carried out using the computer software extraction was carried out using a standard Chelex Gene Profiler (Scanalytics). extraction protocol using a 96-well format microtitre Statistical analysis. The power of the 3 microsatellite plate (information available from the authors upon markers to detect multiple paternity was assessed request). through simulation analysis as implemented in the Microsatellite primer development for both Munida Probability to Detect Multiple Matings (PRDM) pro- rugosa and M. sarsi followed the protocol described by gram (Neff & Pitcher 2002). This simulation takes into Kijas et al. (1994) based on enriched partial genomic consideration the potential number of sires, degree of libraries, with modifications as reported by Boston paternal skew and brood size analysed under a range et al. (2009) and McInerney et al. (2008, 2009). of possible mating scenarios across both species. Mat- Microsatellite-containing regions in the genome of ing scenarios, which take into consideration equal, both M. rugosa and M. sarsi are particularly complex, moderately skewed and highly skewed male breeding being characterised by the presence of cryptic success, were chosen to reflect both information avail- repeated elements including transposable elements able from crustacean mating system studies (Walker et (Bailie et al. 2010). This makes the development of al. 2002, Bilodeau et al. 2005, Yue et al. 2010) and rec- microsatellites for these species exceptionally difficult. ommendations by Neff & Pitcher (2002). Allelic fre- Nevertheless, following the approach outlined by quency population data for Munida rugosa and M. Bailie et al. (2010), 3 informative microsatellite markers sarsi samples from the same area were available from were successfully developed for M. rugosa (see Table 1 a parallel study into the population structure of both for details). While no microsatellite markers were suc- species, which was carried out within our research cessfully developed for M. sarsi, all 3 of the M. rugosa group. markers were found to cross-hybridise and were used The occurrence of multiple paternity of a brood was to screen all samples. established by the occurrence of more than 2 paternal Single locus polymerase chain reaction (PCR) amplifi- alleles across at least 2 loci (to allow for the possibility cations for microsatellite genotyping in a Li-Cor system of mutation at 1 locus). On confirmation of multiple were carried out in 12 µl reaction volumes containing 1× paternity, GERUD (Jones 2001) was used to estimate

Promega Taq polymerase buffer, 1.5 to 2.5 mM MgCl2 the minimum number of possible males, to reconstruct 100 µM dNTP, 1 to 2 pM of each microsatellite primer, all possible male genotypes and to rank the likelihood 50 ng template DNA and 0.5 to 1 U of Promega Taq DNA of each male’s contribution to the brood in correlation polymerase. PCR cycling conditions consisted of 1 cycle with the known maternal genotypes. When more than at 95°C for 5 min, followed by 24 to 28 cycles at 95°C for 2 paternal alleles were identified at a single locus only, 1 min, 42 to 56°C for 1 min, 72°C for 1 min, followed by χ2 statistics were used to test whether the remaining 1 1 cycle of 72°C for 5 min (Table 1). Amplified products or 2 loci displayed evidence for significant deviations were denatured at 80°C for 3 to 5 min, and 1 µl was from expected Mendelian genotypic ratios.

Table 1. Primer details of microsatellite loci developed in this study. Product size represents the size in base pairs (bp) of the cloned allele from which primers were designed, annealing temperature (Ta), MgCl2 concentration and number of PCR cycles used in amplification reactions with this Ta (denaturation and extension temperatures were 95 and 72°C, respectively, for all primer sets). F: forward; R: reverse. *: specific primer of the pair labelled with fluorescent dye

Primer Sequence 5’ to 3’ Product size (bp) Ta (°C) MgCl2 (mM) No. cycles

MR62F TAAACGACCAATCCCATTAGAC 154 52 1.5 28 MR62R* TATATTTGGAGTAAAGTGG MR63F* TCTTGAGAAAGATAGAAATAT 132 42 2.5 25 MR63R CTTGCGCAAGCGGGAATAA MR778F GGAAACCAACTCATTATTACTTAC 135 56 2.5 24 MR778R* CCGTGGCCACCCCCTTAG 176 Mar Ecol Prog Ser 421: 173–182, 2011

RESULTS Table 3. Munida rugosa and M. sarsi. Probability of detecting multiple paternity (PRDM) for 2 or 3 microsatellite loci (MR: M. rugosa, no. of individuals, n = 3; MS: M. sarsi, n = 2) used assuming 3 mating scenarios: (1) even contribution, Although very difficult to develop, (2) moderately skewed towards 1 male and (3) largely skewed towards 1 male the resulting 3 microsatellite marker (or 2 males when 4 were taken into consideration). This was conducted against loci (MR62, MR63 and MR778) were 2 to 4 males and with ranging brood sizes (N). Average numbers of M. rugosa found to be sufficiently informative for and M. sarsi eggs screened were 32 and 40, respectively. Bold type refers to values discussed in text. No. males = mating scenario, Ratio MC = Ratio of parentage analysis in both Munida male contributions rugosa and M. sarsi. Summary genetic statistics for the population samples No. Ratio Species Brood sizes (N) screened for both species (M. rugosa: males MC 10 20 32 40 60 86 no. of individuals, n = 80 and M. sarsi: n = 18 from an ongoing population study) are reported in Table 2. MR778 2 50:50 MR 0.831 0.886 0.893 0.895 0.894 0.894 MS 0.863 0.918 0.925 0.924 0.923 0.925 was found to be monomorphic in M. 67:33 MR 0.794 0.876 0.893 0.894 0.892 0.896 sarsi; thus, the screening of M. sarsi MS 0.829 0.908 0.923 0.924 0.925 0.925 broods was conducted with only 2 95:5 MR 0.272 0.454 0.607 0.668 0.770 0.834 microsatellite loci (MR62 and MR63). MS 0.289 0.486 0.634 0.698 0.801 0.866 No significant departure from Hardy- 3 33:33:33 MR 0.946 0.983 0.988 0.988 0.988 0.988 Weinberg equilibrium was observed MS 0.963 0.992 0.995 0.994 0.995 0.995 globally in either species. No geno- 57:28.5:14.5 MR 0.905 0.971 0.983 0.985 0.987 0.988 MS 0.927 0.982 0.991 0.993 0.994 0.995 typic disequilibrium was observed be- 86:9:5 MR 0.593 0.808 0.905 0.938 0.967 0.980 tween loci, and no null alleles were MS 0.620 0.837 0.928 0.955 0.980 0.990 detected in the present study. 4 25:25:25:25 MR 0.976 0.996 0.998 0.998 0.999 0.999 While only 2 (Munida sarsi) or 3 (M. MS 0.985 0.999 1.000 1.000 1.000 1.000 rugosa) microsatellite markers were 40:20:20:20 MR 0.966 0.994 0.998 0.998 0.998 0.999 used for parentage analysis, these in- MS 0.979 0.998 0.999 1.000 1.000 1.000 variably displayed adequate power (27 40:40:10:10 MR 0.945 0.987 0.995 0.996 0.998 0.998 to 100%) for detecting multiple pater- MS 0.962 0.994 0.998 0.999 0.999 1.000 nity in both species depending on the mating scenario considered (Table 3). Not surprisingly, the power to detect multiple paternity The average number of eggs (32 for Munida rugosa increases with brood size. The logistical difficulties in and 40 for M. sarsi) for which consistent genotypic data characterising genotypes of the high number produced were obtained per brood was adequate for detecting by some crustacean species has resulted in studies (e.g. multiple paternity with a PRDM between 60 and 100% Urbani et al. 1998, Walker et al. 2002, Gosselin et al. across both species (Table 3). Dependent on the skew 2005) implementing a pooling approach. The drawback detected, the average number of eggs (N) screened of such an approach is that the true number of males can does not greatly improve the likelihood of detecting be significantly underestimated, particularly if the male multiple paternity. For example, for M. rugosa with a contribution is highly skewed; therefore, we elected to mating scenario of slight paternal skew (e.g. 67:33), the screen individual eggs. power to detect paternity when screening 20 eggs is 87%, with increased screening of 86 eggs only improv- ing the power to 89%. This indicates that the power Table 2. Munida rugosa and M. sarsi. Summary statistics for of detecting multiple paternity is surprisingly high population samples of M. rugosa (no. of individuals, n = 80) and M. sarsi (n = 18). Allelic diversity (K) observed and despite the low number of markers available. The expected heterozygosities (Hobs and Hexp, respectively) for number of additional paternal alleles observed among 3 microsatellite loci the offspring of M. rugosa broods provided evidence of multiple paternity by, firstly, displaying more than 2

Species Locus KHobs Hexp paternal alleles in more than 1 locus, or when more than 2 paternal alleles were identified at a single locus M. rugosa MR62 14 0.845 0.825 deviating from the Mendelian inheritance of a 1:1 ratio MR63 6 0.436 0.445 (Table 4). MR778 7 0.493 0.496 Multiple paternity was evident in 21 (84%) of the M. sarsi MR62 14 0.960 0.920 MR63 4 0.210 0.270 Munida rugosa families. Thirteen females (52%) MR778 Monomorphic mated with a minimum of 2 sires while the remaining 8 (32%) mated with a minimum of 3 (Table 4). Given Bailie et al.: Multiple paternity in Munida spp. 177

Table 4. Munida rugosa. (A) Allelic inheritance microsatellite DNA profiles for 25 berried females. In each case, allelic profiles, identified by the length of the amplification product (in bp), result from the screening of an average of 32 fertilised eggs. Female alleles and putative sire alleles (in italics) are reported per locus. (B) Average minimum number of males per family with which each female must have mated in order to generate the observed allelic distribution among the offspring is also reported. FG: female genotype

(A) M. rugosa Microsatellite loci Min. female MR62 MR63 MR778 no. ID FG Paternal alleles FG Paternal alleles FG Paternal alleles males

1a,b 204, 232 204, 212, 240 133, 133 133 135, 135 127, 131,135, 139 3 2b 212, 212 204, 212, 220 133, 133 133, 137 123, 131 135 2 3 204, 204 204, 212 133, 133 133 127, 135 135 1 4a,b 204, 204 204, 208, 212, 240 133, 137 133, 137 131, 135 127, 131, 135, 139 3 5a,b 212, 220 212, 220, 208, 216, 232 133, 137 133, 137 135, 135 135 3 6b 204, 240 204, 240, 212 133, 133 133, 137 131, 135 127, 135 2 7a,b 208, 220 200, 204, 208, 216 133, 133 133, 137, 141 135, 135 131, 135 3 8b 196, 208 196, 212, 240 133, 137 133, 137 135, 151 135 2 9a,b 212, 224 196, 204, 212, 228 133, 137 133, 137 131, 139 107,131, 135, 139 3 10 212, 212 206, 208 133, 137 133, 141 135, 139 131, 135 1 11 204, 204 204, 212, 240 133, 137 133, 137 135, 135 131, 135, 139 2 12b 196, 204 204, 212 133, 133 133, 125, 141 135, 135 131, 135 2 13b 204, 212 204, 208 133, 137 141 135, 135 127, 135, 139 2 14a 204, 208 204, 208, 212, 236 133, 133 133 135, 135 127, 135, 139 2 15b 212, 216 204, 212, 220 133, 133 133, 141 131, 139 135, 139 2 16b 204, 212 204, 212 133, 137 133, 137 135, 135 131, 135, 139 2 17b 204, 236 208, 212, 216, 133, 133 133, 137 135, 135 135, 139 2 18b 204, 216 204 137. 137 133 131, 135 127, 135, 139 2 19b 212, 240 208, 212, 220, 228, 232 125, 129 133 135, 135 135, 139 3 20a,b 212, 236 200, 204, 208, 212, 220, 236 133, 137 133, 137 135, 139 127, 135, 139 3 21a,b 204, 208 204, 212, 216, 220, 232 133, 133 133 135, 135 127, 135, 139 3 22 212, 240 212 133, 133 133, 145 135, 135 135 1 23a 204, 204 204, 208, 212 133, 133 133, 137 135, 135 127, 131, 135 2 24a 204, 204 204, 208, 212 133, 137 133, 125, 145 135, 135 131, 135, 139 2 25 204, 208 204, 240 133, 137 133, 137 135, 135 127, 135 1

(B) Min. no. sires Percent of total broods No. broods (= mating females)

3328 25213 1164 aFamilies with multiple paternity detected at more than 1 locus or with more than 3 paternal alleles at a single locus; the likelihood of a false result due to mutation is decreased in these cases bFamilies with multiple paternity detected by significant deviations from expected Mendelian genotypic ratios

that evidence for multiple paternity in 44% of the Table 5. Munida rugosa. Summary of the estimated frequency cases (11 females) is provided by more than 1 locus or of multiple paternity. Six scenarios were considered for the number of fathers and their reproductive skew. The probabil- a single locus with more than 2 paternal alleles, it is ity of detecting this mating scenario in the current M. rugosa unlikely that this is an artefact of high levels of dataset was simulated by the Probability to Detect Multiple mutation. Matings (PRDM). Values discussed in the text are in bold In 66.7% of the detected cases of multiple paternity, paternal contribution was often significantly skewed No. Paternal No. Frequency of Probability (t-test, p < 0.001) towards 1 particular male (i.e. sum of fathers skew families observed of detecting 38.2 and 28.5% from mating scenarios 95:5 and 86:8:6, pattern (%) pattern respectively; Table 5). The estimated PDRM probabil- 2 50:50 1 4.7 0.754 ity of detecting more than 1 male within a given family 67:33 3 14.4 0.748 using these markers ranged from 46 to 93% (Table 5). 95:5 8 38.2 0.466 The numbers observed in the present study are within 3 33:33:33 1 4.7 0.931 the upper extreme of this range (i.e. 84% resulting 57:28.5:14.5 2 9.5 0.922 from the sum of 32 and 52% for 8 and 13 families asso- 86:8:6 6 28.5 0.783 ciated with 3 and 2 sires, respectively, see Table 4). 178 Mar Ecol Prog Ser 421: 173–182, 2011

Father 3 Father 2 Father 1 Unresolved Father 4 Father 3 Father 2 Father 1 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30

% clutch sired 30 % clutch sired 20 20 10 10 0 0 1234 123 Pleopodal region Pleopodal region

Fig. 1. Munida rugosa. Paternal contributions made to Fam- Fig. 2. Munida sarsi. Paternal contributions made to Family 5 ily 7 female’s clutch based upon the highest likelihood ranked female’s clutch based upon the highest likelihood ranked combination of paternal males through GERUD analysis. combination of paternal males through GERUD analysis. In In this case, contribution of presumed males (represented this case, contribution of presumed males (represented by by varying shades of grey) is given by pleopodal region. varying shades of grey) is given by pleopodal region. Four Three sires were identified and contributed to each of the sires were identified in this family, and eggs unable to be 4 pleopodal segments assigned to an individual male due to the allelic makeup of the offspring and other putative males were classified as ‘unresolved’ Conversely, for the remaining 25% of cases, the mark- ers provide little power to detect multiple paternity, hence potentially explaining the instances of single (unless very underrepresented) also contributed to 2 to paternity observed (identified in 16% of cases, see 3 of the 4 pleopodal segments in 52% of broods (e.g. Table 4). The probability of detecting multiple pater- Fig. 1), thereby suggesting the possibility of egg mix- nity decreases (e.g. 46 to 78% when taking into ing or simultaneous egg extrusion. account 2 to 3 males, respectively) with increasing In congruence with Munida rugosa, the allelic distri- paternal skew (Table 5). However, despite this and bution in all 5 M. sarsi broods (100%) provided extreme paternal skew (i.e. >95%), multiple paternity unequivocal evidence of multiple paternity. A mini- was detected in two-thirds (66.7%) of Munida rugosa mum of 4 sires were required to explain each brood’s families in this investigation. allelic make up (Table 6). This contrasts with the The microspatial distribution of contributing males results observed for M. rugosa, where a maximum of 3 in Munida rugosa broods exhibiting multiple paternity sires was detected. Similar to what was observed in (86%) was evident in each of the 4 pleopodal seg- M. rugosa, there appears to be considerable variation ments. It was also clear that second and third males in the relative contribution of individual males siring

Table 6. Munida sarsi. Summary of allelic inheritance from a sample of an average of 40 brooded offspring from each of 5 berried females. Female alleles and male alleles (in italics) are reported (excluding fertilised eggs for which parentage could not be unambiguously resolved among putative males). Each female must have mated with a minimum of 4 males in order to generate the observed allelic distribution among the offspring (as estimated using GERUD). All families had multiple paternity detected at more than 1 locus or with more than 3 paternal alleles at a single locus, and the likelihood of a false result due to mutation is decreased in these cases. In addition, all families had multiple paternity detected by significant deviations from expected Mendelian genotypic ratios

M. sarsi Microsatellite loci female ID MR62 MR63 Female genotype Paternal alleles Female genotype Paternal alleles

1 192, 212 188, 192, 200, 212, 216, 224, 232 129, 133 129, 145 2 188, 224 148, 168, 172, 188, 196,200, 224, 284 129, 129 117, 129, 141 3 212, 220 188, 192, 200, 212, 216, 220, 236 129, 129 129, 133, 141 4 220, 224 188, 212, 220, 224, 244, 308 129, 133 117, 129, 133, 145 5 200, 328 200, 236, 316, 324 129, 129 129, 137, 141 Bailie et al.: Multiple paternity in Munida spp. 179

the offspring of particular females. A single male was typus (Thiel & Hinojosa 2003), porcelain crab Petrolis- highly favoured in all mating scenarios, but the vary- thes cinctipes (Toonen 2004), snow crab Chionoecetes ing proportions of contributions made by subsequent opilio (Urbani et al. 1998) and crayfish Orconectes males differed. As for M. rugosa, when more than 1 placidus (Walker et al. 2002). Contrary to these studies, male was present (in all 5 families) each male’s contri- however, the incidence of polyandry in both Munida bution was evident in each of the 3 pleopodal seg- rugosa (84%), but in particular M. sarsi (100%), is ments (Fig. 2), with few exceptions. higher than that observed in other crustacean species. Prior to the present study, the highest levels of multiple paternity have been observed in species such as O. DISCUSSION placidus (60%, Walker et al. 2002), porcelain crab (80%, Toonen 2004) and Norway lobster Nephrops Given the difficulties in obtaining microsatellites for norvegicus (54%, Streiff et al. 2004). these and possibly other galatheid species (Bailie et al. Previous reports of multiple paternity in crustaceans 2010), those identified in the present study are not only have elaborated on a number of explanations for this extremely valuable but are currently the only markers behaviour, including convenience polyandry, cryptic available for these galatheids. In the present study, female choice (by re-mating, discarding sperm from these 3 markers were found to be sufficiently polymor- selected males or delaying ovulation) or forced copula- phic to detect genetic contribution from multiple tions (Thiel & Hinojosa 2003, Walker et al. 2002). Alter- males. Invariably, while more than 1 marker is needed natively, the reduced number of males resulting from to discard mutation bias, the detection of multiple sex-biased fisheries may result in a female mating with paternity does not require a large number of marker more than 1 male in order to obtain sufficient sperm to loci. Thus, the number of markers used in this case is fertilise her brood (Gosselin et al. 2005). The multiple comparative to the majority of other crustacean molec- mating identified in Munida rugosa and M. sarsi ular paternity studies published to date (e.g. Urbani et females is compatible with 1 or more of these hypothe- al. 1998, Walker et al. 2002, Streiff et al. 2004, Toonen ses. Limited laboratory observations have shown that 2004, Bilodeau et al. 2005, Gosselin et al. 2005). The forced copulations occur in M. sarsi (Pothanikat 2005), secretive and nocturnal nature of galatheids (De Grave while sex-biased fisheries have also been reported for & Turner 1997) makes behavioural observations diffi- M. rugosa (Coombes 2002). cult. Molecular work has proven essential to provide Ra’anan & Sagi (1985) suggested that forced copula- novel information on the mating strategy of these elu- tions, where males transfer sperm externally to the sive galatheids. Indeed, the present study provides the females, are common in crustacean species. This argu- first genetic evidence for the occurrence of multiple ment is corroborated by a number of observations of paternity in any Galatheidae species. this behaviour in species such as the crayfish Aus- In addition to identifying multiple paternity, this is tropotamobius italicus (Rubolini et al. 2006), rock the first investigation to demonstrate that polyandry is shrimp (Thiel & Hinojosa 2003) and Munida sarsi the predominant mating strategy in Munida rugosa (Pothanikat 2005). It is interesting to note that for spe- and M. sarsi from the Clyde Sea area. We argue that cies where forced copulations—involving external the few instances of monogamy identified in M. rugosa sperm transfer and external fertilisation—occur, multi- (14%) are likely to be an artefact of the comparatively ple mating has often been observed (Walker et al. low power of markers employed in this particular case. 2002, Thiel & Hinojosa 2003). It is possible that forced Thus, the frequency of multiple paternity observed is copulations may be triggered by a number of factors. likely to be an underestimation of the true extent for For instance, males might have to force copulation this species in the wild. The identification of a higher because hard-shelled females do not require protec- number of contributing males to broods in M. sarsi in tion, and hence can potentially afford to be choosy. comparison with that seen in M. rugosa broods is inter- Alternatively, following Bateman’s principle, males esting. While we only examined 5 M. sarsi broods, this might want to increase their biological fitness by mat- species occurs at higher densities than M. rugosa ing with a large number of females. (Hartnoll et al. 1992), making it feasible for M. sarsi to While the incidence of forced copulations by males have more opportunity for male encounters. of Munida species implies strong male influence on The findings of this investigation (i.e. polyandry) are mating behaviour, female choice cannot be underesti- comparable to what has been observed in the major- mated. For the rock shrimp, based on behavioural ity of other crustacean species including the thalassi- observations, Thiel & Hinojosa (2003) suggested that nidean ghost shrimp Callichirus islagrande (Bilodeau cryptic female choice plays an important part in this et al. 2005), American lobster Homarus americanus species’ mating behaviour. The authors reported that (Gosselin et al. 2005), rock shrimp Rhynchocinetes Rhynchocinetes typus actively chooses male sper- 180 Mar Ecol Prog Ser 421: 173–182, 2011

matophores favouring larger individuals. Thus, at least Munida species in that area as compared to that seen for this species, females can influence the breeding in unexploited areas. success of particular males even though copulations While external fertilisation may be followed by were forced. Therefore, for rock shrimp species, direct either simultaneous or consecutive egg extrusion in genetic benefits might be the main factor favouring correspondence to each spermatophore, it is currently this particular mating strategy (reviewed by Jennions not known whether galatheid females can handle & Petrie 2000). more than 1 spermatophore at any given time. Consec- In order to counteract male bias through forced cop- utive spermatophore handling and egg extrusion has ulations and to reduce the risk of injury, it is possible been reported in the rock shrimp (Thiel & Hinojosa that females are able to make postcopulatory choices 2003), whereas a simultaneous mode of handling has by selecting spermatophores from large males. This been recorded in the mole crab Emeritus asiaticus mechanism aids to increase genetic diversity and via- (Subramoniam 1977, 1979). If females are capable of bility of the females’ offspring, while also ensuring the handling more than 1 spermatophore as eggs are provision of an adequate amount of sperm for the fer- being extruded, then males may compete by removing tilisation of their broods. While sperm competition other males’ sperm (Beninger et al. 1991). Thus, and/or sperm pack removal by male competitors can- despite theories of cryptic female choice, the observed not be excluded, it is also possible that the paternal paternal skew could alternatively be the result of last skew in the parentage assignment observed in Munida male precedence, as has previously been reported in species is due to female cryptic choice. Therefore, it is the snow crab (Sévigny & Sainte-Marie 1996). There- feasible that females could potentially favour ‘better’ fore, the identification of significantly highly skewed males. broods (1 male sired >86% of the brood; 66% of The already established squat lobster fisheries in Munida rugosa families and 100% of M. sarsi families) western Scotland and reported squat lobster discards could be the result of cryptic female choice and/or last through Nephrops fisheries (Bergmann & Moore 2001) male precedence and/or sperm competition. have the potential to perturb the mating system of the For both Munida species reported in the present species. Direct or indirect fishing efforts may affect the study, the offspring from different sires were randomly operational sex ratio (OSR) of the Munida species, distributed among the females’ brood-patch. While which can influence behavioural aspects of each sex. such an observation is compatible with a simultaneous This is particularly true when one sex is gathered in fertilisation mode, the alternative hypothesis of higher numbers than the other, causing a skew in the sequential egg extrusion cannot be ruled out if egg sex ratio, thus increasing competition for each respec- mixing occurs in the brood. Considering that egg mix- tive sex. ing may result in egg loss, it is reasonable to assume Studies of population biology conducted by that simultaneous fertilisation is the most likely sce- Coombes (2002) suggested that male Munida rugosa nario for these species. were dominant in particular months of the year (i.e. February and June), and another study found com- paratively reduced numbers of females in May and CONCLUSION August (Hartnoll et al. 1992). Although it is unlikely that this reflects a true difference in sex ratio, the The genetic parentage analyses conducted in this skew may represent cryptic behavioural responses study have demonstrated that (1) both Munida rugosa related to the reproductive and moulting cycle, and M. sarsi mate multiply with 2 to 4 sires, as respec- resulting in males being more vulnerable to fishing tively 86 and 100% of broods were multiply sired; pressures in particular months. The incidence of mul- (2) the offspring of different males were distributed tiple paternity may be influenced by geographical throughout the females’ pleopods; and (3) in most location or fisheries’ exploitation levels as suggested broods analysed (66 and 100%, respectively), it was for the American lobster (Gosselin et al. 2005). The evident that they were significantly highly skewed removal of large males by fisheries forces females to towards a single male. In this particular study, we can- multiply mate with smaller males in order to get not differentiate between all the theories discussed enough sperm to fertilise their broods. Males have (convenience polyandry as a result of cryptic female also been reported to economise their sperm (MacDi- choice, forced copulations and/or the influence of fish- armid & Butler 1999, Rondeau & Sainte-Marie 2001, ing pressures) or a combination thereof. Future work Rubolini et al. 2005). This behaviour also forces would benefit from comparisons of mating scenarios females to mate with more than 1 male to ensure fer- from fisheries exploited and unexploited areas and tilisation of their broods. Consequently, the Clyde Sea more detailed behavioural observations to disentangle fisheries may affect the mating system observed in the possible theories discussed. Bailie et al.: Multiple paternity in Munida spp. 181

Acknowledgements. This work was funded by a grant from Jennions MD, Petrie M (2000) Why do females mate multiply? the Department of Agriculture and Rural Development A review of the genetic benefits. Biol Rev Camb Philos Soc (DARD), Northern Ireland. Many thanks to T. Claverie from 75:21–64 University of Glasgow for providing squat lobster samples Jivoff P, Hines AH (1998) Female behaviour, sexual competi- from western Scotland. Crustacean genetics research in tion and mate guarding in the blue crab, Callinectes P.A.P.’s laboratory is currently supported by the Beaufort sapidus. Anim Behav 55:589–603 Marine Research Award in Fish Population Genetics carried Jones A (2001) GERUD 1.0: a computer program for the out under the Sea Change Strategy and the Strategy for Sci- reconstruction of parental genotypes from progeny arrays ence Technology and Innovation (2006-2013), with the sup- using multilocus DNA data. Mol Ecol Notes 1:215–218 port of the Marine Institute, funded under the Marine Kijas JMH, Fowler JCS, Gabett CA, Thomas MR (1994) Research Sub-Programme of the National Development Plan Enrichment of microsatellites from the citrus genome 2007–2013. using biotinylated oligonucleotide sequences bound to streptavidin-coated magnetic particles. Biotechniques 16:657–662 LITERATURE CITED Lebour M (1930) The larvae of the Plymouth Galatheidae. I. Munida banffica, Galathea strigosa and Galathea dis- Bailie DA (2008) Phylogeny, population genetics and mating persa. J Mar Biol Assoc UK 17:175–188 strategies of the squat lobster species Munida rugosa and MacDiarmid AB, Butler MJ (1999) Sperm economy and limi- Munida sarsi (Crustacea, Decapoda, Galatheidae). PhD tation in spiny lobsters. Behav Ecol Sociobiol 46:14–24 thesis, Queen’s University, Belfast McInerney CE, Allcock AL, Johnston MP, Prodöhl PA (2008) Bailie DA, Fletcher H, Prodöhl PA (2010) High incidence of Characterisation of polymorphic microsatellites for the cryptic repeated elements in microsatellite flanking periwinkle gastropod, Littorina littorea (Linnaeus, 1758) regions of galatheid genomes and its practical implica- and their cross-species amplification in four congeners. tions for molecular marker development. J Crustac Biol 30: Conserv Genet 10:1417–1420 664–672 McInerney CE, Allcock AL, Johnson MP, Prodöhl PA (2009) Beninger PG, Elner RW, Poussarty Y (1991) Gonopods of the Characterization of polymorphic microsatellites for the majid crab Chionoecetes opilio (O. Fabricius). J Crustac rough periwinkle gastropod, Littorina saxatilis (Olivi, Biol 11:217–223 1792) and their cross-amplification in four congeners. Bergmann M, Moore PG (2001) Survival of decapod crus- Conserv Genet 10:1989–1992 taceans discarded in the Nephrops fishery of the Clyde McKeown NJ, Shaw PW (2008) Polymorphic nuclear Sea area, Scotland. ICES J Mar Sci 58:163–171 microsatellite loci for studies of brown crab, Cancer pagu- Bilodeau AL, Felder DL, Neigel JE (2005) Multiple paternity rus L. Mol Ecol Resour 8:653–655 in the thalassinidean ghost shrimp, Callichirus islagrande Neff BD, Pitcher TE (2002) Assessing the statistical power of (Crustacea: Decapoda: Callianassidae). Mar Biol 146: genetic analyses to detect multiple mating in fish. J Fish 381–385 Biol 61:739–750 Boston ESM, Montgomery WI, Prodöhl PA (2009) Develop- Pothanikat RME (2005) The behaviour and ecology of Munida ment and characterisation of 11 polymorphic compound rugosa and Munida sarsi. PhD dissertation. Queen’s Uni- tetranucleotide microsatellite loci for the Leisler’s bat, versity, Belfast Nyctalus leisleri (Vespertilionidae, Chiroptera). Conserv Ra’anan Z, Sagi A (1985) Alternative mating strategies in Genet 10:1501–1504 male morphotypes of the freshwater prawn Macro- Coombes JCH (2002) Aspects of the biology and fisheries brachium rosenbergii (De Man). Biol Bull (Woods Hole) ecology of the velvet swimming crab, Necora puber (L.), 169:592–601 and the squat lobsters Munida rugosa (Fabricius) and M. Rondeau A, Sainte-Marie B (2001) Variable mate guarding sarsi Huss (Crustacea: Decapoda) in Scottish waters. PhD time and sperm allocation by male snow crabs (Chinoe- dissertation, University of Glasgow cetes opilio) in response to sexual competition, and their Correa C, Thiel M (2003) Mating systems in caridean shrimp impact on the mating success of females. Biol Bull (Woods (Decapoda: Caridae) and their evolutionary consequences Hole) 201:204–217 for sexual dimorphism and reproductive biology. Rev Chil Rubolini D, Galeotti P, Ferrari G (2006) Sperm allocation in Hist Nat 76:187–203 relation to male traits, female size, and copulation behav- De Grave S, Turner JT (1997) Activity rhythms of the squat iour in a freshwater crayfish. Behav Ecol Sociobiol 60: lobsters, Galathea squamifera and G. strigosa (Crustacea, 212–219 Decapoda, Anomura) in south-west Ireland. J Mar Biol Sainte-Marie B, Sevigny JM, Gauthier Y (1997) Laboratory Assoc UK 77:273–276 behaviour of adolescent and adult males of the snow crab Duffy JE, Thiel M (eds) (2007) Evolutionary ecology of social (Chionoecetes opilio) (Brachyura: Majidae) mated non- and sexual systems: crustaceans as model organisms. competitively and competitively with primiparous Oxford University Press, New York, NY females. Can J Fish Aquat Sci 54:239–248 Gore RH (1979) Larval development of Galathea rostrata Schmidt WL (1965) Crustaceans. The University of Michigan under laboratory conditions, with a discussion of larval Press, Ann Arbor, MI development in the Galatheidae (Crustacea Anomura). Sévigny JM, Sainte-Marie B (1996) Electrophoretic data sup- Fish Bull 76:781–806 port the last-male sperm precedence hypothesis in the Gosselin T, Sainte-Marie B, Bernatchez L (2005) Geographic snow crab, Chionoecetes opilio (Brachyura: Majidae). variation of multiple paternity in the American lobster J Shellfish Res 15:437–440 Homarus americanus. Mol Ecol 14:1517–1525 Streiff R, Mira S, Castro M, Cancela ML (2004) Multiple pater- Hartnoll RG, Rice AL, Attrill MJ (1992) Aspects of the biology nity in Norway lobster (Nephrops norvegicus L.) assessed of the galatheid genus Munida (Crustacea, Decapoda) with microsatellite markers. Mar Biotechnol 6:60–66 from the Porcupine Seabight, Northeast Atlantic. Sarsia Subramoniam T (1977) Aspects of sexual biology of the 76:231–246 anomuran crab, Emerita asiatica. Mar Biol 43:369–377 182 Mar Ecol Prog Ser 421: 173–182, 2011

Subramoniam T (1979) Some aspects of reproductive ecology Van Dover CL, Williams AB (1991) Egg size in squat lobsters of a mole crab Emerita asiatica (Milne Edwards). J Exp (Galatheoidea): constraint and freedom. In: Wenner A, Mar Biol Ecol 36:259–268 Kuris A (eds) Crustacean egg production. Crustacean Taggart JB, Hynes RA, Prodohl PA, Ferguson A (1992) A sim- issues, Vol 7. Balkema, Rotterdam, p 143–156 plified protocol for routine total DNA isolation from Wada S, Ashidate M, Goshima S (1997) Observations on the salmonid fishes. J Fish Biol 40:963–965 reproductive behaviour of the spiny king crab Paralith- Tapella F, Lovrich GA, Romero MC, Thatje S (2002) Repro- odes brevipes (Anomura: Lithodidae). Crustac Res 26: ductive biology of the crab Munida subrugosa (Decapoda: 56–61 Anomura: Galatheidae) in the Beagle Channel, Argentina. Walker D, Porter BA, Avise JC (2002) Genetic parentage J Mar Biol Assoc UK 82:589–595 assessment in the crayfish Orconectes placidus, a high Thiel M, Hinojosa IA (2003) Mating behaviour of the female fecundity invertebrate with extended maternal brood rock shrimp Rhynchocinetes typus (Decapoda: Caridea)— care. Mol Ecol 11:2115–2122 indication for convenience polyandry and cryptic female Wenner EL (1982) Notes on the distribution and biology of choice. Behav Ecol Sociobiol 55:113–121 Galatheidae and Chirostylidae (Decapoda: Anomura) Toonen RJ (2004) Genetic evidence of multiple paternity of from the middle Atlantic Bight. J Crustac Biol 2:360–377 broods in the intertidal crab, Petrolisthes cinctipes. Mar Yue GH, Li JL, Wang CM, Xia JH, Wang GL, Feng JB (2010) Ecol Prog Ser 270:259–263 High prevalence of multiple paternity in the invasive Urbani N, Sainte-Marie B, Sevigny JM, Zadworny D, Kuhn- crayfish species Procambarus clarkii. Int J Biol Sci 6: lein U (1998) Sperm competition and paternity assurance 107–115 during the first breeding period of female snow crab (Chi- Zainal KAY (1990) Aspects of the biology of the squat lobster noecetes opilio) (Brachyura: Majidae). Can J Fish Aquat Munida rugosa (Fabricius, 1775). PhD dissertation, Uni- Sci 55:1104–1113 versity of Glasgow

Editorial responsibility: Karen Miller, Submitted: July 2, 2010; Accepted: October 26, 2010 Hobart, Australia Proofs received from author(s): January 5, 2011